- 29 Jun, 2006 29 commits
-
-
Cornelia Huck authored
Fixes for several channel measurement facility bugs: * Blocks copied from the hardware might not be consistent. Solve this by moving the copying into idle state and repeating the copying. * avg_sample_interval changed with every read, even though no new block was available. Solve this by storing a timestamp when the last new block was received. * Several locking issues. * Measurements were not reenabled after a disconnected device became available again. * Remove #defines for ioctls that were never implemented. Signed-off-by:
Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by:
Martin Schwidefsky <schwidefsky@de.ibm.com>
-
Gerald Schaefer authored
Add CPU ID and steal time, and make OS record size variable. Signed-off-by:
Gerald Schaefer <geraldsc@de.ibm.com> Signed-off-by:
-
Peter Oberparleiter authored
Implementation of new kernel parameter vmpanic that provides a means to perform a z/VM CP command after a kernel panic occurred. Signed-off-by:
Peter Oberparleiter <peter.oberparleiter@de.ibm.com> Signed-off-by:
-
Horst Hummel authored
Add support for parallel-access-volumes to the dasd driver. This allows concurrent access to dasd devices with multiple channel programs. Signed-off-by:
Horst Hummel <horst.hummel@de.ibm.com> Signed-off-by:
-
Martin Schwidefsky authored
Remove export of the sys_call_table symbol to prevent the misuse of it. Signed-off-by: -
Martin Schwidefsky authored
The two macros NEW_TO_OLD_UID and NEW_TO_OLD_GID in binfmt_elf32.c are not used by any code. Remove them. Signed-off-by: -
Serge E. Hallyn authored
overflowuid and overflowgid were exported twice. Remove the export from s390_ksyms.c Signed-off-by:
Serge E. Hallyn <serue@us.ibm.com> Signed-off-by:
-
Cornelia Huck authored
After setting a path to a dasd offline at the SE, I/O hangs on that dasd for 5 minutes, then continues. I/O for which an interrupt will not be reported after the channel path has been disabled was not terminated by the common I/O layer, causing the dasd MIH to hit after 5 minutes. Be more aggressive in terminating I/O after setting a channel path offline. Also make sure to generate a fake irb if the device driver issues an I/O request after being notified of the killed I/O and clear residual information from the irb before trying to start the delayed verification. Signed-off-by:
-
Gerald Schaefer authored
Remove EXPORT_SYMBOL_GPL(avenrun) from appdata_base.c, since it is already exported in kernel/timer.c Signed-off-by:
-
Serge E. Hallyn authored
as a deprecated kernel_thread to a kthread. Signed-off-by:
-
Stefan Weinhuber authored
The struct dasd_eer_header needs the packed attribute, or there will be 6 additional bytes of random data between the fixed header and the variable length part of the eer data. Signed-off-by:
Stefan Weinhuber <wein@de.ibm.com> Signed-off-by:
-
Heiko Carstens authored
The preempt_count in the thread_info structure must be initialized to 1. Signed-off-by:
Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by:
-
Heiko Carstens authored
There is almost no room left for any new code between 0x10000 and 0x10480. Move the code from 0x10000 to 0x11000. Signed-off-by:
-
Horst Hummel authored
Dasd code cleanup: 1) remove white space, 2) remove the emacs override sections, and 3) use kzalloc instead of kmalloc. Signed-off-by:
-
Martin Schwidefsky authored
If a machine checks interrupts the external or the i/o interrupt handler before they have completed the cpu time calculations, the accounting goes wrong. After the cpu returned from the machine check handler to the interrupted interrupt handler, a negative cpu time delta can occur. If the accumulated cpu time in lowcore is small enough this value can get negative as well. The next jiffy interrupt will pick up that negative value, shift it by 12 and add the now huge positive value to the cpu time of the process. To solve this the machine check handler is modified not to change any of the timestamps in the lowcore if the machine check interrupted kernel context. Signed-off-by: -
Gerald Schaefer authored
Signed-off-by:
-
Horst Hummel authored
The ccw dump function dasd_eckd_dump_ccw_range can crash because it does not take care about the IDAL flag in the ccw. Check for IDALs flag set in CCW and follow the indirect list to print the data that is refered by the ccw. Signed-off-by:
-
Stefan Weinhuber authored
Check the return value of kzalloc in dasd_eer_open. Signed-off-by:
-
Martin Schwidefsky authored
The initial i/o to a 3270 device is done using the static module variables raw3270_init_data and raw3270_init_request. If the 3270 device driver is built as a module and gets loaded above 2GB, the initial i/o will fail because these variables will get addresses > 2GB. To make it work the two variables are moved to struct raw3270 and the data structure is allocated with GFP_DMA. Signed-off-by: -
Martin Schwidefsky authored
The software watchdog calls machine_restart from a timer function. The s390 machine_restart calls console_unblank to flush the console output. This is needed for panic to get the panic message printed. If console_unblank is called in interrupt a BUG is triggered in acquire_console_sem. That makes the software watchdog panic instead of restarting the machine. To get around this problem the call to console_unblank is made conditionally on !in_interrupt() || oops_in_progress. Signed-off-by: -
Heiko Carstens authored
Fix __syscall_return macro: valid error numbers are in the range of -1..-4095. Signed-off-by:
-
Peter Oberparleiter authored
Changes in the DASD driver require an asynchronous implementation of the subchannel reprobe loop. This loop was so far only used by the blacklisting mechanism but is now available to all CCW device drivers. Signed-off-by:
-
Cornelia Huck authored
Work around the problem that a device cannot be unregistered from driver_for_each_device() because of klist node refcounting: Get device after device owned by the driver to be unregistered with driver_find_device() and then unregister it. This works because driver_get_device() gets us out of the region of the elevated klist node refcount. driver_find_device() will always get the next device in the list after the found one has been unregistered. Signed-off-by:
-
Cornelia Huck authored
Specify correct sizeof() in chp_measurement_read() and return correct amount of read data. Signed-off-by:
-
Peter Oberparleiter authored
Trying to set a DASD root device online can fail under some circumstances with the message "Read configuration data returned error -5". The cause is that read configuration data incorrectly aborts with -EIO when it encounters a temporary busy condition at a storage server. Perform retry when encountering temporary busy conditions. Signed-off-by:
-
Heiko Carstens authored
The wrong base register is used to read a value from the sclp data structure. The value is used to calculate the memory size. Use correct register %r4. Signed-off-by:
-
Heiko Carstens authored
show_stack() passes a pointer to the current stack frame to show_trace(). Because of tail call optimization the pointer doesn't point to the original stack frame anymory and therefore traces are wrong. Don't pass the pointer of the current stack frame to show_trace(). Instead let show_trace() calculate the pointer on its own. Signed-off-by:
-
Heiko Carstens authored
Encapsulate complete bitops.h with #ifdef __KERNEL__ and remove the now superfluous ALIGN_CS define and its users. This patch is needed for compiling klibc. Signed-off-by:
-
Linus Torvalds authored
The recent vsnprintf() fix introduced an off-by-one, and it's now possible to overrun the target buffer by one byte. The "end" pointer points to past the end of the buffer, so if we have to truncate the result, it needs to be done though "end[-1]". [ This is just an alternate and simpler patch to one proposed by Andrew and Jeremy, who actually noticed the problem ] Acked-by:
Andrew Morton <akpm@osdl.org> Acked-by:
Jeremy Fitzhardinge <jeremy@goop.org> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
- 28 Jun, 2006 11 commits
-
-
Linus Torvalds authored
* 'nommu' of master.kernel.org:/home/rmk/linux-2.6-arm: [ARM] nommu: backtrace code must not reference a discarded section [ARM] nommu: Initial uCLinux support for MMU-based CPUs [ARM] nommu: prevent Xscale-based machines being selected [ARM] nommu: export flush_dcache_page() [ARM] nommu: remove fault-armv, mmap and mm-armv files from nommu build [ARM] Remove TABLE_SIZE, and several unused function prototypes [ARM] nommu: Provide a simple flush_dcache_page implementation [ARM] nommu: add arch/arm/Kconfig-nommu to Kconfig files [ARM] nommu: add stubs for ioremap and friends [ARM] nommu: avoid selecting TLB and CPU specific copy code [ARM] nommu: uaccess tweaks [ARM] nommu: adjust headers for !MMU ARM systems [ARM] nommu: we need the TLS register emulation for nommu mode
-
Linus Torvalds authored
* 'devel' of master.kernel.org:/home/rmk/linux-2.6-arm: [ARM] 3672/1: PXA: don't probe output GPIOs for interrupt [ARM] 3671/1: ep93xx: add cirrus logic edb9315 support [ARM] 3370/2: ep93xx: add crunch support [ARM] 3665/1: crunch: add ptrace support [ARM] 3664/1: crunch: add signal frame save/restore [ARM] 3663/1: fix resource->end off-by-one thinko during physmap conversion [ARM] 3662/1: ixp23xx: don't include asm/hardware.h in uncompress.h [ARM] 3660/1: Remove legacy defines [ARM] 3661/1: S3C2412: Fix compilation if CPU_S3C2410 only [ARM] 3658/1: S3C244X: Change usb-gadget name to s3c2440-usbgadget [ARM] Remove the __arch_* layer from uaccess.h
-
Linus Torvalds authored
* master.kernel.org:/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog: [WATCHDOG] Documentation/watchdog update [WATCHDOG] convert AT91RM9200 watchdog to platform driver [WATCHDOG] add WDIOC_GETTIMELEFT ioctl [WATCHDOG] Pre-Timeout flags
-
Thomas Gleixner authored
plist.h uses container_of, which is defined in kernel.h. Include kernel.h in plist.h as the kernel.h include does not longer happen automatically on all architectures. Signed-off-by:
Thomas Gleixner <tglx@linutronix.de> Signed-off-by:
-
David Brownell authored
This patch adds earlier initialization of spi_device.mode, as needed on boards using nondefault chipselect polarity. An example would be ones using the RS5C348 RTC without an external signal inverter between the RTC chipselect and the SPI controller. Without this mechanism, the first setup() call for that chip would wrongly enable chips, corrupting transfers to/from other chips sharing that SPI bus. Signed-off-by:
David Brownell <dbrownell@users.sourceforge.net> Signed-off-by:
-
Linus Torvalds authored
* master.kernel.org:/pub/scm/linux/kernel/git/dtor/input: Input: db9 - fix potential buffer overrun Input: atkbd - fix hardware autorepeat Input: wistron - add mapping for Wistron MS 2111
-
Greg Ungerer authored
Remove the hard coded ROM region setups. Use Kconfig options to specify these in a generic way for platorms that want them. This builds on top of the other recent m68knommu linker script changes to completely remove fixed board configurations. Signed-off-by:
Greg Ungerer <gerg@uclinux.org> Signed-off-by:
-
Greg Ungerer authored
Use Kconfig options to setup the optional ROM region used on some platforms. We used to define this in the linker script on a per board basis. The configure options are more flexible and clean up the linker script a lot. Signed-off-by:
-
Greg Ungerer authored
Fix 68EZ328/config.c asm to be clean for new gcc versions. Signed-off-by:
-
Greg Ungerer authored
Fix 68VZ328/config.c asm to be clean for new gcc versions. Signed-off-by:
-
Greg Ungerer authored
Fix 68360/config.c asm to be clean for new gcc versions. Signed-off-by:
-
